Supply Chain Data Integration SCDI for Operational Decision Support

1
Supply Chain Data Integration (SCDI) for
Operational Decision Support
Logistics Research Network Annual Conference
2004 September 9 and 10 Quinn School of
Business Dublin Søren Selmar Kenn
Steger-Jensen Denmark
2
Programme

• Introduction
• The MMM-foundation
• SCDI scope of application
• Problems
• The SCDI-challenges
• The data selection problem
• Solutions
• The SCDI system
• Search models and search strategies
• Access privileges in SCDI
• Supporting data integration in four ways
• The virtual ODS
• Supporting evolution and self-optimization
• SCDI in a ASP business model
• Summing up
• Questions Discussion

3
A Data Warehouse Architecture
/Ken Orr, 2000/
4
How to build a Data Warehouse
/Oracle, 2001/
5
The SCDI Challenges

• Distributed databases in upper SME supply chains
• Large numbers of companies and databases can be
  involved
• Dynamic, complex and evolutionary supply chains
• Heterogeneous and unsteady system environment
• Limited IT resources and limited financial
  resources
• Competitors and collaborators? Data security is
  essential!
• User-needs when selecting data for operational DS
• Transparent cross-company data search
• Near real-time data
• Intelligent and interactive search support
• End-user oriented and user-friendly interface
• Support of unpredictable ad hoc search
• Support of exception reports and standard reports

6
The Data Selection Problem
Data Models Mapping information Database
Designs Database Allocation Identification
System SQL or Tools
RDB 1
RDB 2
RDB 3
RDB 3
RDB 3
RDB n
7
Solutions

• Data selection using the SCDI system
• The users models
• Structures supported by SCDI
• The search strategy in SCDI
• Three ways of making data integration in a SCDI
  system
• The virtual ODS
• How to support evolution and self-optimization
• The SCDI-system supported by an ASP

8
The SCDI System
Users
RDB 1
RDB 2
Users Models
Generic Models
RDB 3
RDB 3
RDB 3
RDB n
Business Intelligence systems
9
Users Models Supported by SCDI
10
Generic Models Supported by SCDI

• Organizational structure
• Classification structure. Inheritance
• Decomposition Aggregation structure
• Bill of material
• Activity structure
• Geographical structures
• Any other relevant structure
• supply chain model
• process model
• product model
• etc.

11
The Search Strategy in SCDI (1)

• Any ad hoc search in SCDI can follow the same few
  user-friendly steps
• Step 1. Chose among presented subjects (like
  product)
• Step 2. Chose among presented properties (like
  selling price)
• Step 3. Put on data constraints (like selling
  price gt 300 )

12
The Search Strategy in SCDI (2)

• More can be added to specify a data selection
• Step 00. Chose or modify existing search
  specification
• Step 0. Organizational limitation of the search
• Step 1. Chose among presented subjects (like
  product)
• Step 2. Chose among presented properties (like
  selling price)
• Step 3. Put on data constraints (like selling
  price gt 300 )
• Step 4. Chose appropriate output format
• Step 5. Output data documentation and statistics
  (Yes/No)
• Step 6. Specify search service levels (like 80
  and 20 sec.)
• Step 7. Save as a standard report or as an
  exception report

13
Access Privileges in SCDI

• Every company connected to the SCDI-system decide
  who is going to have access, and to what part of
  their data
• Access privileges are made almost like specifying
  a data selection
• It means that an access privilege can be
  specified by
• Subject(s) (Step 1)
• Properties (Step 2)
• Data constraints (Step 3)
• Company, Server, Database, Table and Column
  (Special step)
• In the data selection interface only metadata
  corresponding to the personal access privileges
  are shown
• Access privileges can be inherited from an upper
  organizational unit or from another person

14
Four Ways of Making Data Integration in the SCDI
System

• Virtual ODS (Operational Data Store)
• Real ODS. Data replicated. Special design
• Real ODS. Data replicated. One-to-one design
• DWH. Data replicated. Search-optimized design

15
How does SCDI Support Evolution and
Self-optimization?

• Full independency between the interface layer and
  the source database designs
• Dynamic and differentiated change of the SCDI
  database design can be utilized 4 levels are
  defined
• The optimizer component must support
• Definition of service levels ( , sec.)
• Monitoring service levels
• Optimizing the database designs (4 levels)
• Establish replication and update the metabase

16
The Evolutionary SCDI System Design
17
SCDI in a ASP Business Model.Pros and Cons

• Pros
• Effective use of system resources
• Effective use of SCDI competences
• Secure handling of access-privileges
• Customer defined service levels
• Cons
• How to trust an ASP
• A SCDI-success can be expensive .

18
Summing up

• Problems
• The SCDI-challenges in general
• The Data Selection problem
• Solutions
• The data selection method
• The virtual ODS
• The evolutionary and self-optimizing ODS
• SCDI support by a ASP (Application Service
  Provider)
• Recommendation
• Start with internal use of the SCDI-method
• Improve data quality and data documentation
• Improve IT agility and readiness for real supply
  chain data integration

19
A SCDI System Makes Data Integration Attractive
and Affordable for the SMEs
Users
RDB 1
RDB 2
Users Models
Generic Models
RDB 3
RDB 3
RDB 3
RDB n
Business Intelligence systems
20
Questions and Discussion

• Thank You for Your Attention